Changelog

Material changes to the Luminid platform and methodology. For complete methodology version history, see the methodology page.

May 2026

Skill taxonomy rebuilt on an open standard. Our skill library now sits on a single, standards-based taxonomy. Every verification maps to a shared vocabulary, so a skill you prove means the same thing everywhere on Luminid and stays portable wherever you take it.

Groundwork for evidence-cited scoring. We started building the layer that ties every verification to the specific work that earned it, so a verified skill always points back to demonstrated evidence you can check.

Third blog post published: "Removing AI inference from experience scoring." Why methodology 2.0 removed experience and education from the Role Match formula, and how it exemplifies the way AI is constrained at Luminid. Read the post

Second blog post published: "What construct validity means in hiring." The framework that separates assessment from theater, and how Luminid earns construct validity through ongoing calibration rather than vendor claims. Read the post

About page published. The story behind Luminid, the founder background, and the Costa Rica origin. The founder voice companion to the methodology page. Read about Luminid

Methodology page v2.0 published. The complete Luminid methodology is now public, covering the Role Match formula, AI scope and constraints, bias-handling architecture, calibration loop, and known limitations. Read the methodology

First blog post published. "Why we don't score resumes" covers the Schmidt and Hunter meta-analysis, the Luminid decision to exclude experience and education from the Role Match, and the "verified hiring" category framing.

Methodology 2.0 shipped. Experience, education, and language self-declaration removed from the Role Match. Candidates are now evaluated on demonstrated skills and simulation performance only. Experience and education appear on candidate profiles for recruiter context but do not drive the Role Match. Methodology page v2.0